Engineering Study Project
Name
Florides , MF
Login Identifier
a72253483
Email Address
[email protected]
Course/Stage
EEE (H605)/3 Silicon Nanowires and their
Project Title
Applications
Supervisor
Prof. A.G. O’Neil
Submission Date
25/01/10
Copyright © 2010, Michalis Florides
Silicon nanowires and their applications Abstract This paper describes the characteristics of silicon nanowires (SiNW) and their use in sensor and transistor applications based on existing literature. An introductory part talks about nanowires in general and how they can affect the evolution of next generation electronics and electrochemical biosensors. A reference to the properties of silicon nanowires describes their mechanical, electrical, chemical, optical and thermal properties. Furthermore, a description of the way nanowire based sensors are implemented and their characteristics is done plus factors which affect their performance and possible applications. The implementation of silicon nanowire based transistors is discussed in addition with their characteristics and related applications. Finally, this paper concludes with the most important thinks regarding silicon nanowires.
1. Introduction Nanowires can be defined as nanostructures which have diameters towards the nanometer scale (in the area of tens of nanometers). It is possible to make conducting (e.g. Ni, Pt, Au), semiconducting (e.g. Si, InP, GaN) and insulating (e.g. SiO2, TiO2) nanowires which can find different possible real applications [16]. The basic methods for the synthesis of nanowires are the bottom-up method and the top-down method which is similar to lithography [1, 16]. The bottom-up method is based on chemical composition and is the preferred method for the nanowires’ synthesis because important parameters of a nanowire such as radius, length, doping levels, chemical composition and growth direction can be controlled accurately [4, 14, 16]. Nanowires are still in an experimental stage and they do not have use in real applications [16]. Recent investigations suggested that they can be used as the building blocks for the next generation electronics and very sensitive biosensors [2, 12, 16]. Another possible real application of nanowires is the nano electromechanical systems (NEMS) due to their high Young’s moduli [4, 16,]. Nanorobots can also be aided by nanowires in the generation and conduction of their required energy [17]. Silicon nanowires are considered as popular nanomaterials due to their exceptional electrical and mechanical properties [5]. They are semiconducting nanowires and their conductivity can be controlled by the field-effect action [3, 16, 18]. For this reason, silicon
1 Copyright © 2010, Michalis Florides
nanowires can be used probably for the next generation of field-effect transistors and advanced sensors [7, 9]. Moore’s law states that the number of transistors on an integrated circuit will double approximately every two years [4, 17]. Until now Moore’s Law holds true, however, further down scaling will be required by the next generation electronics. This is because the power consumption has to be reduced and at the same time the performance of the devices has to be improved. The device integration density has to be increased too. The International Technology Roadmap of Semiconductors (ITRS) expects that the metal oxide semiconductor field-effect transistors (MOSFETs) will shrink down to 18nm and their gate length will be scaled down to 7nm. Thin film bulk MOSFETs are not expected to meet these requirements and thus new structures and fabrication methods have to be found if Moore’s Law will continue to be true [4]. Silicon nanowires are candidates for the new generation electronics because of their exceptional properties and their comparable performance with the best planar devices [4, 5, 6, 18]. Medicine and biochemistry will probably require new advanced sensors for the diagnosis of diseases and the discovery of new drugs. These sensors are required to have fast real-time response, low cost and high resolution of analysis [8]. What is more, sensitivity down to a single molecule is desirable plus direct label-free detection [13, 18]. The extraordinary properties and structural characteristics of silicon nanowires promise to give rise to these new types of biosensors and electrochemical sensors and meet the desired requirements [9].
2. Properties 2.1 Mechanical properties Mechanical properties of nanowires are of great importance in device processing since changes in temperature, induced strain and external stress can change the electrical conductivity of the nanowire due to internal dislocations. The compressive and tensile stresses during VLSI (Very Large Scale Integration) processing can cause failure due to delamination and electro-migration [4]. It is expected that nanowires, which are 1-D systems, have interesting mechanical properties due to their high aspect ratio compared to bulk materials and the reduced number of defects per unit length [4, 12]. However, manipulating these materials for mechanical measurements is a challenging task [4, 6]. The main methods which are used to investigate their mechanical properties are mechanical resonant, the Atomic Force Microscope (AFM) and
2 Copyright © 2010, Michalis Florides
nanoindentation [5, 6]. The resonant method is used only for the determination of the elastic properties and it is not easy to measure the applied force accurately when using the AFM method. On the other hand, nanoindentation has very good force and displacement resolution and control [6]. Experiments, based on the AFM-based nanoindentation, showed that the stiffness of silicon nanowires is well described by the Herz theory and that nanowires with diameters between 100nm and 600nm have elastic modulus values which are more or less the same as the values for bulk silicon. This implies that the elastic characteristics of silicon nanowires with diameters of 100nm are not affected by the finite size effect. What is more, the elastic modulus of nanowires with diameters less than 100nm was found to reduce with decreasing diameter [5]. Nanowires can be used in the field of sensors and nano-electromechanical systems (NEMS) [4, 5]. This is because, according to their tensile strength and Youngs Modulus, they are very robust materials and have the ability to store elastic energy [4, 11]. In addition, nanoscale resonators can be built by silicon nanowires with high oscillating frequencies (100 MHz up to 1 GHz) due to their excellent elastic properties [4, 5]. The detection of molecules at atomic resolution was achieved with these nanoscale resonators [5].
2.2 Electrical properties It is important to understand the electrical properties of semiconducting nanowires because they determine the suitability of silicon nanowires to be used in electronics and sensor applications. Silicon nanowires have four atoms per unit cell. If the crystalline is perfect, three conductance channels are found. Conductance variations are observed if one or two atoms are added or removed. Thus the conductance is affected by the crystalline structure of the nanowire [9]. What is more, variations in the surface conditions, like scattering phenomena of carriers in nanowires, cause changes in conductivity [4, 9]. Scattering phenomena are observed when the diameters of the nanowires are changed [9]. The large surface-to-volume ratio of nanowires makes their conductivity very sensitive to surface excitation by external charges [9, 11, 12]. As it is seen in figure 2.2, a small surface perturbation can influence the ‘bulk’ of a nanowire while in the case of a planar device only a fraction of its surface is influenced. This important property allows the detection of a single molecule and the use of silicon nanowires in biosensors [7, 8, 11, 18].
3 Copyright © 2010, Michalis Florides
Figure 2.2: The major advantage of 1-D nanostructures (B) over 2-D thin film (A). Binding to 1-D nanowire leads to depletion or accumulation in the “bulk” of the nanowire as opposed to only the surface in 2-D thin-film case [11].
In addition, lateral current shunting is caused in 2-D structures. This is avoided in nanowires due to their 1-D structure, and thus signal intensities are not reduced. This property allows for a direct electrical readout when the nanowire is used as a sensor or a field effect transistor and makes real-time detection possible [7, 11]. Furthermore, the free carrier concentration in silicon nanowires depends on the size of the structure. It was shown that the donor ionization energy of silicon nanowires increases when the radius of the nanowire decreases. Therefore, the free carrier density can be modified at radius values much larger than those at which quantum and dopant segregation effects set in [3]. Finally, in addition to the surface sensitivity of nanowires, other unique properties such as quantum confinement effects and low leakage currents make nanowires attractive for the next generation of electronics and nanosystems [4, 9].
2.3 Chemical properties An important chemical reaction, which makes silicon nanowires very useful in sensor and transistor applications, is the natural oxidation. This is an unavoidable effect which creates oxygen-derived defects on the surface of a pure semiconductor nanowire core. It was shown, by experimental measurements, that the hole mobility of a silicon nanowire can be increased by a magnitude of two if there is proper chemical modification of the SiOx layer [1].
2.4 Optical properties Nanowires can exhibit mechanical strain effects if exposed to light which has a
4 Copyright © 2010, Michalis Florides
wavelength comparable to their energy bandgap. This is due to their photoelastic properties [5].
2.5 Thermal properties Silicon nanowires, when used within applications or experiments, may have a curved like shape and not be straight. The phonon transport can be affected by their curvature and thus their thermal conductivity changes. The nanowire curvature is a mean of impedance to the phonon transport because phonons deviate from the main heat flow direction which flows axially through the nanowire. Therefore, thermal conductivity is reduced when the radius of the nanowire curvature decreases [2]. The effect of the curvature on the thermal impedance has greater effect when the radius of the curvature is one order smaller than the phonon mean-free path (MFP) [2, 15]. This observation is interesting since the thermal conductivity of silicon nanowires can be controlled by proper shaping of the wire [2]. This is important in the use of silicon nanowires in next generation electronics because the shrinking of electronic devices towards the nanoscale region demands an increase in power dissipation per unit area [15].
3. Sensor applications Sensors are important tools for life sciences and biochemistry. The use of sensors in these areas leads to the detection and diagnosis of diseases and to the discovering of new drugs [8, 18]. Microelectronic sensors based on thin film transistors and ion-sensitive filed-effect transistors (ISFETs) were used until the 1970s. They offered a cheap solution over the chemical sensors and could be integrated on a chip [8]. Nevertheless, the microelectronic sensors did not have the required characteristics to be used as biosensors of their undesired sensitivity to temperature and light and because their parameters were not fixed over the time. Also, solidstate electrodes were not reliable and this led to the use of chemical sensors [8]. The detection of biomolecules at low concentrations is achieved with fluorescent labeling and optical detection methods. However, this technique is expensive and time-consuming and thus silicon nanowires may be an alternative solution [7, 8].
3.1 Implementation and operation Sensors are composed of an active sensing element and a signal transducer. The active sensing element has to catalyze the material being sensed and produce an output signal which can be electrical, thermal or optical [9].
5 Copyright © 2010, Michalis Florides
A silicon nanowire can be connected between two electrodes as shown in figure 3.1. to form a sensor. The surface of the SiNW is covered with a receptor (active element) which can bind only to the desired molecule to be detected. Once this specific molecule reaches the surface of the SiNW its conductivity (transducer) increases or decreases depending on the type of the nanowire (p-type or n-type) and on the charge of the molecule [18]. Also, the electrodes are protected by an oxide layer to ensure that the only change in the SiNW conductivity is due to process occurring only near the surface of the nanowire [8, 18].
Figure 3.1: Schematic of a Si nanowire-based FET device configured as a sensor with antibody receptors (green), where binding of a protein with net positive charge (red) yields a decrease in the conductance [18].
The example in figure 3.1 shows a p-type doped silicon nanowire (p-Si) covered with antibody receptors. It is seen that when a protein molecule which has a net positive charge in an aqueous solution reaches the surface of the SiNW its conductance drops [18]. This action is similar to a FET transistor where the conductance of its channel can be controlled by the voltage applied to its gate [13, 18]. In the case of a SiNW sensor there is not a gate electrode and the action of the gate voltage is done by the charge of the molecule being sensed [7, 18].
3.2 Characteristics-Performance Silicon nanowires can almost act as perfect biosensors due to their inherent properties [7]. When used as sensors their characteristics include ultrahigh sensitivity because the molecule being sensed depletes or accumulates the charges in the bulk of the nanowire [7, 18]. In addition, direct label-free detection allows the molecules to be detected in real time which eliminates the time consuming labeling chemistry [8, 9, 18]. Another important characteristic is that they are non-radioactive and that sensor arrays can be built if the nanowires are
6 Copyright © 2010, Michalis Florides
manufactured with the top-down technology which allows detection of different molecules in the same solution [8, 9]. Although, SiNW sensors have great characteristics there are factors which can affect their performance. These include the surrounding environment of the nanowire and the electrostatic screening action of the ions in the solution. The performance of a sensor can be characterized by the parameters of sensitivity, settling time and selectivity. A simulation with a silicon nanowire with a length between 2μm and 20μm showed how the sensitivity can be influenced by several factors. The results showed that the sensitivity increases with reduced dopant density and decreased diameter and length. However, it is not possible to reduce these quantities as much as needed to achieve the maximum desired sensitivity due to the dopant fluctuations effect. Furthermore, this simulation revealed that the dielectric constant of the surrounding environment affects the sensitivity of SiNW. If the surrounding media is air then the sensor has greater sensitivity if it is designed to operate in the depletion mode. If the surrounding media is water the mode of operation (accumulation/depletion) is not very important as the doping density because of the high dielectric constant of water. Finally, parasitic ions in the surrounding solution of the sensor screen the charge of the target molecule and reduce sensitivity [8].
3.3 Applications Experimental applications of SiNW sensors include the detection of proteins, DNA, pH, drug discovery, single viruses, glucose, arrays for parallel molecule detection and gas sensors [11, 18, 19].
3.3.1 Proteins The detection of the protein streptavidin is achieved by using the molecule biotin, which is highly selective to this protein, as the binding material (figure 3.3.1.1.A). The results are shown in figure 3.3.1.1. When the protein reaches the surface of the SiNW its conductance is increased very fast to a constant value (figure 3.3.1.1.B). In the absence of the molecule biotin the conductance does not change (figure 3.3.1.1.C) and this shows the high selectivity and sensitivity of SiNW sensors [11, 18, 20].
7 Copyright © 2010, Michalis Florides
Figure 3.3.1.1: Real-time detection of proteins and DNA. (A) Schematic of a biotin-modified Si nanowire and subsequent binding of streptavidin to the modified surface. (B) Plot of conductance versus time for a biotin-modified Si nanowire, where region 1 corresponds to the buffer solution, region 2 corresponds to the addition of 250 nM streptavidin, and region 3 corresponds to pure buffer solution. (C) Conductance versus time for an unmodified Si nanowire, where regions 1 and 2 are the same as in (B) [18].
Figure 3.3.1.2 shows the detection of protein specific antigen (PSA) which was performed by a single chip complementary SiNW sensor. The complementary responses are shown and it is obvious that there is no cross-talk between the n-doped and p-doped SiNWs [11].
Figure 3.3.1.2: Complementary sensing of PSA using p-type (NW1) and n-type (NW2) siliconnanowire devices in the same array. The vertical solid lines correspond to times at which PSA solutions of 1) 0.9 ng/mL, 2) 0.9 ng/mL, 3) 9 pg/mL, 4) 0.9 pg/mL, and 5) 5 ng/mL were connected to the microfluidic channel [11].
8 Copyright © 2010, Michalis Florides
3.3.2 DNA When uncharged peptide nucleic acid (PNA) is used on the surface of a p-type SiNW as the binding material DNA detection is possible (figure 3.3.2.D) [9, 11, 18]. The PNA was designed to bind only to wild type DNA and the response of the SiNW sensor is shown in figure 3.3.2.E. The conductance of the SiNW increases rapidly due to the negative charge of the DNA. The inset of figure 3.3.2.E shows the response of the sensor to DF508 mutation DNA. This shows that the sensor does not have a stable response to the mutant DNA and thus the high selectivity of SiNW sensors is again proved. What is more, figure 3.3.2.F shows the response of two independent silicon nanowire sensors (red and blue). The responses of the two sensors are similar and thus silicon nanowire sensors can be manufactured the same characteristics [18].
Figure 3.3.2: (D) Schematic of a Si nanowire sensor surface modified with PNA receptor before and after duplex formation with target DNA. (E) Si nanowire DNA sensing where the arrow corresponds to the addition of a 60 fM complementary DNA sample and the inset shows the device conductance following addition of 100 fM mutant DNA. (F) Conductance versus DNA concentration, where data points shown in red and blue are obtained from two independent devices [18].
3.3.3 Single virus The detection of a single virus molecule is an important fact for medicine and a challenge for SiNW biosensors [11, 18]. In an experiment the surface of the p-type SiNW sensor was covered by an antibody receptor as shown in figure 3.3.3.A (yellow).
9 Copyright © 2010, Michalis Florides
Figure 3.3.3: (A) Schematic of a single virus binding and unbinding to the surface of a Si nanowire device modified with antibody receptors and the corresponding time-dependent change in conductance. (B) Simultaneous conductance and optical data recorded for a Si nanowire device after the introduction of influenza A solution. The images correspond to the two binding/unbinding events highlighted by time points 1-3 and 4-6 in the conductance data, with the virus appearing as a red dot in the images [18].
It is seen that, when the virus molecule (influenza A) binds/unbinds to the antibody receptor (figure 3.3.3.A), the conductance of the SiNW changes. The fact that only a single virus molecule caused the change in conductance was verified by simultaneous monitoring of fluorescently labeled viruses. The change of conductance of the SiNW is shown in figure 3.3.3.B too. It is obvious that the conductance is influenced only when the virus (red dor) comes in contact with the sensor. This observation suggests that nanowire device arrays can be built in high densities without cross-talk [18].
3.3.4 pH Sensing of pH is actually sensing of the hydrogen ion concentration. A p-type SiNW, when its surface is modified with 3-aminopropyltriethoxysilane, can detect hydrogen ions. The amino and silanol groups act as a catalyst and thus the surface charge of the SiNW changes. As shown in figure 3.3.4.A-B, the conductance of the SiNW varies in a stepwise fashion when the pH of the solution being analyzed is increased from 2 to 9. To underline the importance of the surface receptor for the selectivity of the sensor, measurements of the pH were taken without the 3-aminopropyltriethoxysilane receptor. The response is shown in figure
10 Copyright © 2010, Michalis Florides
3.3.4.D and it is similar to previous measurements derived from silica sensors. What is more, this experiment showed that the sensing mechanism is due to field-effect action [18].
Figure 3.3.4: Nanowire pH sensors. (A) Schematic of an amino-functionalized nanowire device and the protonation/deprotonation equilibria that change the surface charge state with pH. (B) Changes in nanowire conductance as the pH of solutions delivered to the sensor is varied from 2 to 9; inset is a plot of conductance data versus pH. (C) Schematic of an unmodified nanowire sensor containing silanol groups and the protonation/deprotonation equilibria that change the surface charge state with pH. (D) Conductance of an unmodified Si nanowire device (red) versus pH. The dashed green curve is a plot of the surface charge density for silanol groups on silica as a function of pH [18].
3.3.5 Glucose Glucose detection have been reported too. The SiNW surface oxide layer is modified with glucose oxidase (GOx) enzyme. A change in the SiNW conductance was observed when glucose was near the surface of the nanowire [11].
3.3.6 Drug discovery New drugs discoveries are essential for the development of pharmaceuticals. The process is the same as with the previous application examples however the change in the SiNW conduction is used to evaluate how good is the bind of the organic molecule to the
11 Copyright © 2010, Michalis Florides
protein or enzyme (possible receptors). Therefore, SiNW sensors allow for the characterization of drugs easily and quickly [18, 20].
3.3.7 Sensor arrays Silicon nanowires can be integrated into arrays which can be useful for the analysis of complex solutions. The reproducibility of SiNW elements in the array is essential for the performance of the whole sensor [18]. SiNW have good reproducibility as mentioned in section 3.3.2.
Figure 3.3.7.1: (A) Schematic of multiplexed single-virus detection using Si nanowire devices modified with antibody receptors for specific viruses. The specific binding/unbinding and corresponding conductance changes are illustrated for two events involving viruses with opposite charges. (B) Simultaneous conductance versus time data from two Si nanowires elements, where NW 2 was modified with antibodies for influenza A (red data) and NW 1 was modified with antibodies for adenovirus (blue data). Black arrows 1-4 correspond to the introduction of adenovirus, influenza A, pure buffer, and a 1:1 mixture of adenovirus and influenza A. Small red and blue arrows highlight conductance changes corresponding to the diffusion of viral particles past the nanowire and not specific binding [18].
12 Copyright © 2010, Michalis Florides
The idea of a simple array sensor is shown in figure 3.3.7.1.A. The receptor on the surface of SiNW1 (blue) is such that it binds with adenovirus and the receptor on the surface of SiNW2 (red) is such that it binds with influenza A. Both SiNW are p-type and thus, when the adenovirus is delivered which is negatively charged, SiNW1 (blue) shows a positive change in its conductance. SiNW2 (red) does not respond to the presence of the adenovirus because it is targeting the influenza A. On the other hand, when the influenza A is delivered, the conductance of SiNW2 (red) reduces because influenza A has a positive net charge. Also, SiNW1 (blue) is not affected by the influenza A because it is targeting the adenovirus. To ensure that each sensor responds only to its target molecule both viruses are delivered to the sensor at point 4 as shown in figure 3.3.7.1.B. This shows that each sensor can respond independently to its target molecule and is not affected by other molecules in the solution to be analyzed which again verifies the high selectivity of SiNW sensors [18]. Another experiment with a p-type SiNW array sensor was designed to detect three different cancer biomarkers (protein specific antigen-PSA, carcinoembryonic antigenCEA and mucin-1). The sensor arrangement is shown in figure 3.3.7.2.A [11, 20].
Figure 3.3.7.2: (a) Schematic illustration of multiplexed detection of cancer biomarkers. Three parallel, independent p-type Si NW FETs, NW1, NW2, and NW3 were functionalized with anti-PSA, anti-CEA, and anti-mucin-1, respectively. Multiplexed data are shown in (b). Solutions of analytes were delivered to the NW array sequentially as follows: (1) 0.9 ng/ml PSA; (2) 1.4 pg/ml PSA; (3) 0.2 ng/ml CEA; (4) 2 pg/ml CEA; (5) 0.5 ng/ml mucin-1; and (6) 5 pg/mL mucin-1. Clearly, only the device designed to respond to the analyte injected showed a response; the other two showed no response. Buffer solutions were injected following each protein solution at points indicated by black arrows [20].
13 Copyright © 2010, Michalis Florides
Regarding figure 3.3.7.2.B, NW1 targets the PSA, NW2 the CEA and NW3 the mucin-1. The sequence the three cancer biomarkers were delivered is PSA, then CEA and finally mucin-1. The response of each of the sensors is shown in figure 3.3.7.2.B. It is clear that each SiNW sensor responds only to its target molecule. This shows again the potential of integration of SiNW sensors into arrays for the analysis of complex biological solutions [20].
3.3.8 Gas sensors Silicon nanowires can find applications as gas sensors too. This was demonstrated when the surface of an n-type SiNW was coated with palladium nanoparticles (Pd particles). This sensor targets H2 gas molecules and thus it showed no response when it was exposed to NH3 or N2O gases. However, when H2 gas flowed over the sensor, the current flowing through it was increased as shown in figure 3.3.8. Thus, SiNW sensors, when used as chemical sensors, have high selectivity values as when used as biosensors. Furthermore, the SiNW sensor responded faster (2.3sec) than an ordinary macroscopic Pd wire sensor (more than 10s). Note that, the sensor was inside a chamber with a pressure of 0.01 Torr and a voltage of 2V was applied across it [19].
Figure 3.3.8: Real time current response of Pd coated SiNWs in 5% H2. The inset images show the enlarged current response (upper) and a SEM image (below) of the device.
4. Transistor applications Silicon transistor technology, especially MOS technology, is scaling down as predicted by Moore’s Law [4]. Initially, transistors had a 3-D active region (bulk silicon substrate) which reduced to 2-D (ultrathin-film silicon-on-insulator (SOI)) as shown in figure 4.1. Recently, the transistor active region was scaled to 1-D (FinFET on SOI) [12].
14 Copyright © 2010, Michalis Florides
Figure 4.1: Continued scaling of silicon CMOS transistor into nanometer regime requires the corresponding reduction of device active layer dimensionality [12].
However, this scaling of the typical silicon transistor technology has almost reached its limits [4, 7]. Despite the advanced fabrication tool capabilities, many physical effects can prevent the tools from working satisfactorily as the devices’ size is scaled down to the nanometer scale. In addition, transistor scalability, performance and power dissipation are three basic problems the top-down technology faces. Furthermore, leakage mechanisms associated with the scalability of conventional silicon transistors down to the nanoscale include gate dielectric direct tunneling, band to band tunneling and short channel effects. Finally, silicon has almost reached its intrinsic switching speed limit. Other semiconducting materials with higher carrier mobility could solve the problem but the required power dissipation levels in the nanoscaled structure does not allow further down scaling [12]. One dimension structures (1-D) are the smallest structures which can be used for efficient transport of electrons [10, 11, 12]. Semiconducting nanowires belong to this category and they are candidates to replace the ultrathin-film SOI transistors. The reasons are that they can be manufactured in large quantities with almost identical characteristics as required by VLSI systems and because they can be manufactured using the bottom-up method rather than the top-down [14]. Bottom-up method advantages over the top-down method include well controlled device dimensions and channel width, very low power dissipation, high integration density, cheap self-assembly technology and enhanced carrier mobility due to reduced scattering (more perfect structure) and even electrical integrity [12, 14]. Silicon nanowires have been studied more and are referred as the most suitable for the implementation of nanowire transistors because silicon dominates the semiconductor industry and because their structure and doping can be controlled accurately [14].
15 Copyright © 2010, Michalis Florides
4.1 Silicon nanowire transistor implementation-operation A simple silicon nanowire transistor can be built as shown in figure 4.1.1. The two gold cubes labeled with S and D (figure 4.1.1.A) represent the source (S) and drain (D) contacts of the transistor respectively. The dark blue cylinder represents the nanowire which is the channel of the transistor. The channel can be doped p-type or n-type. The gate of the transistor is shown in figures 4.1.1.B-C as a rectangular like plate labeled with G. The gate can be placed in a semi-cylindrical shape on the top of the nanowire or all around the the nanowire [14].
Figure 4.1.1: Schematic of NWFETS with (a) back gate, (b) semicylindrical top gate, and (c) cylindrical gate-all-around configurations. The nanowire is dark blue, gate-dielectric is light purple, and source (S), drain (D), and top-gate (G) electrodes are gold. Insets show device cross section at midpoint between source and drain [14].
The operation of this SiNW transistor is similar to that of a typical FET transistor (figure 4.1.2). If the SiNW is p-type and a positive/negative voltage is applied to the gate (G) then the carriers are depleted/accumulated. On the other hand, if the SiNW is n-type and positive/negative voltage is applied to the gate then the carriers are accumulated/depleted. Therefore, the variation of the SiNW conductance via the field-effect action allows for the transistor action to be implemented with SiNWs [13, 18].
Figure 4.1.2: (a) 20 nm p-Si NW array device (channel length of 1 μm; from red to pink, Vg = -5 V to 3 V); and (b) 20 nm n-Si NW device (channel length of 2 μm; from yellow to red, Vg = -5 V to 5 V) in a standard back-gated NWFET geometry as illustrated. Insets in (a) and (b) are current versus gate-voltage (Ids-Vg) curves recorded for NWFETs plotted on linear (blue) and log (red) scales at Vds = -1 V and 1 V, respectively [13].
16 Copyright © 2010, Michalis Florides
Another possible way to fabricate a SiNW based transistor is the one shown in figure 4.1.2. This is the twin SiNW FET which is actually a Gate-All-Around MOSFET (GAA MOSFET) built with SiNWs. The twin SiNW FET shows very good subthreshold characteristics and the gate lengths can be reduced down to 30nm [7].
Figure 4.1.2: Twin SiNW fabrication steps [7]
4.2 Characteristics As was seen in section 4.1, SiNW transistors can be manufactured with a gate which surrounds the whole nanowire (all-around-gate). This allows for better current control through the channel and thus higher current densities can be controlled compared to a planar device. What is more, the all-around-gate allows the devices to be shrunk even more (down to 10nm) due to the excellent control of short-channel effects and leakage [4]. P-type SiNW have attracted greater interest than n-type ones. P-type SiNWs were fabricated and showed high performance characteristics. Their transconductance was about 10 times greater than typical planar devices and the holes mobility was an order of magnitude larger too. Both the transconductance and mobility increased after surface modification of the SiNW. This implies that some of the electrical characteristics of the SiNW transistors can be controlled by proper surface modification. Therefore, the performance of SiNW transistors can exceed that of typical devices [4]. On the other hand, the contact resistance of SiNW FETs affects their performance significantly. Devices with titanium (Ti) source-drain contacts revealed variations in their transconductance and mobility after thermal annealing effects. Other materials for the contacts, like silicide and nickel monosilicide, can reduce this variation in the performance of the device [4].
17 Copyright © 2010, Michalis Florides
4.3 Applications SiNW transistors can find numerous real applications such as displays, data storage, 3-D computing, lasers, smart cards, wearable electronics, high efficiency programming, ring oscillators etc [12, 13, 14]. Some of them are summarized below.
4.3.1 Memory array The memory array shown in figure 4.3.1 is based on vertically grown SiNWs. The unique array architecture and small cell size may provide high density memory at low fabrication cost with fast programming or accessing capability. Also, the speed to density ratio is higher while the cost to density ratio is lower [12].
Figure 4.3.1: (a) Vertically oriented 1-D semiconductor nanowire array potentially utilized to build high-density memory chip. (b) Breaking the barrier of speed–density trade-off. (c) Cost-effectiveness of the vertical nanowire memory technology. [12]
4.3.2 Crossed nanowire architecture Crossed nanowire architecture allows a variety of electronic device elements to be built (transistors, diodes, gates etc) with a high integration density due to the tiny diameter of SiNWs. ‘Crossed NWFETs can be configured from one NW as the active channel and the second crossed NW as the gate electrode separated by a thin SiO2 dielectric shell on the SiNW surface, with the gate on the surface of one or both of the crossed NWs’ [13].
18 Copyright © 2010, Michalis Florides
The first example of the crossed nanowire architecture is the creation of a NOR gate as shown in figure 4.3.2.A. It is shown that the output of the gate goes high only when its input are both low as expected. Another example is the implementation of a 4-bit address decoder as shown in figure 4.3.2.B. Its output has a complementary response with respect to its input (figure 4.3.2.C).
Figure 4.3.2: Crossed NW electronic devices. (left) Schematic of a logic NOR gate constructed from a one-by-three crossed NW junction array using one SiNW and three GaN NWs; insets show a representative scanning electron micrograph of the device (scale bar, 1 μm) and symbolic electronic circuit. (right) Output voltage versus the four possible logic address level inputs; inset is the Vo-Vi relation, where the solid and dashed red (blue) lines correspond to Vo-Vi1 and Vo-Vi2 when the other input is 0 (1). (b) Schematic and scanning electron micrograph of a four-by-four crossed Si NW array address decoder, with four horizontal NWs (I1 to I4) as inputs and four vertical NWs (O1 to O4) as signal outputs. The four diagonal cross points in the array were chemically modified (green rectangles) to differentiate their responses from to the input gate lines. Scale bar, 1 μm. (c) Real-time monitoring of the Vg inputs (blue) and signal outputs (red) for the four-by-four decoder [13].
4.3.3 Ring oscillator Another experimental application of SiNW transistors is the ring oscillator (figure 4.3.3). One of its advantages over conventional fabrication methods is that the whole device integration is achieved during fabrication because of the high reproducibility of
19 Copyright © 2010, Michalis Florides
SiNWs [13]. Also, the great stability and high performance of the SiNW ring oscillator makes it superior over the ones built with conventional planar devices [13, 14].
Figure 4.3.3: (a) Schematic of a multi-NW inverter on a glass substrate and a circuit diagram for a NW ring oscillator consisting of three inverters in series. (b) Oscillation at 11.7 MHz for a p-Si NW ring oscillator with Vsupply = 43 V [13].
5. Conclusions Silicon nanowires are still in an experimental stage but they can be used for the evolution of next generation nanostructures due to their exraordinaty properties compared with other nanomaterials. Their high Young’s moduli and elasticity allows for the development of nano electromechanical systems (NEMS) and flexible electronics. The electrical properties of SiNWs can be well controlled during their synthesis by using the bottom-up fabrication method and in combination with their high reproducibility SiNWs are candidates for the next generation electronics and electrochemical biosensors. Experimental transistor structures made of SiNWs showed that they have the potential to exceed the performance of conventional planar devices. Also, the reported possibility of controlling the thermal conductivity of SiNWs offers higher power dissipation levels per unit area. In addition, the high reproducibility and packing density of SiNW FETs allows for a variety of inexpensive and reliable electronic device elements to be built by using the crossed nanowire architecture. What is more, experimental SiNW chemical and biosensors showed characteristics of ultra-sensitivity and high selectivity. These characteristics can allow detection down to a single molecule level which can have significant impact on the diagnosis of diseases and lead to new drug discoveries. Furthermore, SiNW electrochemical sensors provide direct label-free electrical readout which eliminates the use of the expensive and time consuming chemical labeling detection. Finally, the possible integration of SiNW sensors into arrays allows for fast and accurate analysis of complex chemical/biological solutions.
20 Copyright © 2010, Michalis Florides
6. References 1. Fagas, G. and Greer, JC (2009) ‘Ballistic conductance in oxidized Si’, Nanoletters, 9(5), pp.1856-1860 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 03/12/09) 2. Liu, L.C.; Huang, M.J.; Yang, R.; Jeng, M.S. and Yang, C.C. (2009) ‘Curvature effect on the phonon thermal conductivity of dielectric nanowires’, Journal of Applied Physics, 105(10), Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 03/12/09) 3. Bjork, M.T.; Schmid, H.; Knoch, J.; Riel, H. and Riess, W. (2009) ‘Donor deactivation in silicon nanostructures’, Nature Nanotechnology, 4(2), pp.103-107 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 04/12/09) 4. Wu, X.; Kulkarni, J.S.; Collins, G.; Petkov, N.; Almecija, D.; Boland, J.J.; Erts, D. and Holmes, J.D. (2008) ‘Synthesis and Electrical and Mechanical Properties of Silicon and Germanium Nanowires’, Progress in Chemistry, 20(19), pp.5954-5967 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 04/12/09) 5. Sohn, Y.S.; Park, J.; Yoon, G.; Song, J.; Jee, S.W.; Lee, J.H.; Na, S.; Kwon, T. and Eom, K. (2009) ‘Mechanical Properties of Silicon Nanowires’, Nanoscale Research Letters, pp.1-6 Scopus [Online]. Available at: www.scopus.com (Accessed: 07/12/09) 6. Feng, G; Nix, W.D. and Lee, C.J. (2006) ‘A study of the mechanical properties of nanowires using nanoidentation’, Journal of Applied Physics, 99(7), Scopus [Online]. Available at: www.scopus.com (Accessed: 07/12/09) 7. Najam, S.F (2008) Silicon Nanowire: Synthesis and Applications. Available at: http://swot.ecs.soton.ac.uk/137/3/IRP_edited.pdf (Accessed: 07/12/09) 8. Nair, P.R. and Alam, M.A. (2007) ‘Design Considerations of Silicon Nanowire Biosensors’, IEEE Transactions on Electron Devices, 54(12), pp.3400-3408 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 07/12/09) 9. Yogeswaran, U. and Chen, S.M. (2008) A Review on the Electrochemical Sensors and Biosensors
Composed
of
Nanowires
as
Sensing
Material.
Available
at:
http://www.mdpi.com/1424-8220/8/1/290/pdf (Accessed: 09/12/09) 10. Hu, J.; Odom, T.W. and Lieber, C.M. (2009) ‘Chemistry and Physics in One Dimension: Synthesis and Properties of Nanowires and Nanotubes’, Accounts of Chemical Research, 32(5), pp.435-445 Scopus [Online]. Available at: www.scopus.com (Accessed: 09/12/09)
21 Copyright © 2010, Michalis Florides
11. Wanekaya, A.K.; Chen, W.; Myung, N.V. and Mulchandani, A. (2006) ‘Nanowire-Based Electrochemical Biosensors’, Electroanalysis, 18(6), pp.533-550 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 09/12/09) 12. Yu, B. and Meyyappan, M. (2006) ‘Nanotechnology-Role in emerging nanoelectronics’, Solid-State Electronics, 50(4), pp.536-544 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 10/12/09) 13. Li, Y.; Qian, F.; Xiang, J. and Lieber, C.M. (2006) ‘Nanowire electronic and optoelectronic devices’, Materials Today, 9(10), pp.18-27 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 10/12/09) 14. Lu, W.; Xie, P. and Lieber, C.M. (2008) ‘Nanowire Transistor Performance Limits and Applications’, IEEE Transactions on Electron Devices, 55(11), pp.2859-2876 Web of Knowledge [Online]. Available at: http://apps.isiknowledge.com (Accessed: 10/12/09) 15. Cahay M. (2001) Quantum confinement VI: nanostructured materials and devices. Available
at:
http://books.google.co.uk/books?id=LeGtX_wI6SsC&lpg=PA71&ots=8BFpyRa_CS&d q=phonon%20MFP&pg=PP1#v=onepage&q=phonon%20MFP&f=false
(Accessed:
20/01/10) 16. Wikipedia (2009) Nanowire. Available at: http://en.wikipedia.org/wiki/Nanowire (Accessed: 20/01/10) 17. Strickland
J.
(2010)
How
Nanowires
Work.
Available
at:
http://science.howstuffworks.com/nanowire.htm (Accessed: 22/01/10) 18. Patolsky F. and Lieber C.M. (2005) ‘Nanowire nanosensors’, Materials Today, 8(4), pp.20-28 Scopus [Online]. Available at: www.scopus.com (Accessed: 03/12/09) 19. Chen Z.H.; Jie J.S.; Luo L.B.; Wang H.; Lee C.S. and Lee S.T. (2007) ‘Applications of silicon nanowires functionalized with palladium nanoparticles in hydrogen sensors’, Nanotechnology, 18(34) Scopus [Online]. Available at: www.scopus.com (Accessed: 20/01/10) 20. Curreli M.; Zhang R.; Ishikawa F.N.; Chang H.K; Cote C.R.; Zhou C. and Thomson M.E (2008) ‘Real-time, label-free detection of biological entities using nanowire-based FETs’, IEEE Transactions on Nanotechnology, 7(6) Scopus [Online]. Available at: www.scopus.com (Accessed: 21/01/10)
22 Copyright © 2010, Michalis Florides
